This proposal focuses on the measurement and nature of modifiability in high and low language ability Latino American children. There is evidence in the literature that suggests that observation of modifiability, or ability to change given intervention, can help to distinguish between children who are language different and children with language disorder. This research will first focus on baseline information and validation studies that will guide later studies. The first study explores Latino American children's use of verbal display of knowledge in naturalistic interaction. It is known that Latino American children tend to score low on standardized tests. This is often attributed to differences in home interaction, specifically in demands for verbal display of knowledge. This initial study seeks to confirm differences in home/school demands to predict standardized test performance. The next studies focus on validation of the proposed modifiability scales that will be applied in the final studies using classical test theory. Validation of these measures is critical to ensure reliability and construct validity of the measures. The subjects used in these studies include Latino, bilingual preschoolers ranging in language ability and experiences. The final two years will focus on application of growth curve modeling to predict short term language change and language ability as correlated to modifiability. Studies will employ both qualitative and quantitative methods to examine "moments of learning" which may predict modifiability. If dynamic assessment is to have clinical utility, it is important that methods of exploring change in individual children be examined. The application of modifiability to the problem of providing non-biased assessment options has widespread implications for language assessment of children from culturally/linguistically diverse backgrounds. Increased understanding of the nature and measurement of modifiability may prove to be useful in limiting bias in language assessment and in developing measures of language ability.